[License-discuss] proposal to revise and slightly reorganize the OSI licensing pages

[Chris Travers]

On Thu, Jun 7, 2012 at 8:18 PM, John Cowan <cowan at mercury.ccil.org> wrote:
> Rick Moen scripsit:
>
>> I keep hearing a limited group of people speaking of this alleged tort
>> ('purporting to sublicense'), but fail to find it in copyright law.
>
> Is there actually such a thing as copyright sublicensing?  I suspect not.
> In which case "purporting to sublicense" an unchanged copy of a work
> is usurping the copyright owner's right to control the license, and
> likewise for a copy whose changes are de minimis.  You can license your
> derivative work however you like, consistently with the original license,
> but that's not a sublicense: it is the license of the new work.

Maybe I misunderstood what Larry Rosen was saying about the
differences in the BSD and MIT licenses in his book then ;-).  I
thought his discussion was pretty clear though.  Also see Gardner v.
Nike, 9th Circuit 2002 (IANAL  btw as I will repeatedly state below).
I am not 100% sure but I think after the changes in 2010, exclusive
licensees are now assumed to have sublicense rights as well.  For
non-exclusive licensees (all open source licenses), that's a different
issue.  Maybe a lawyer could correct me if I am wrong about the 2010
changes.

The thing that makes these issues hard is that protecting software
with copyright is a bit like pounding nails with an adjustable wrench.
 The tool isn't really designed for that (copyright, at least in the
US, is designed to protect literature, not recipes in cookbooks) and
so it seems to me there are all sorts of gotchas.

If I give a book publisher the right to sublicense my book, I would
assume at a minimum they could tell a magazine they could serialize
it, for example, and on what terms.  Presumably they could license an
excerpt to be published in an anthology and set terms (within certain
limits dependent on the contract with the publisher) for that
publication.  Maybe they could even negotiate movie rights.  My
understanding is that US law assumes that sublicensing is not allowed
unless specifically stated even in the case of an exclusive copyright
license.  IANAL though.

The Nusphere case is more interesting when we stop thinking about
software and look at copyright as protecting what might be thought of
as "software as literature" or "software as expression."  The GPL
allows mere aggregation without license contagion but requires that
works "based on" the original work carry the same license.  If we
assume that these tie directly to categories of US copyright works,
then "based on" means derivative work (in the sense that a movie might
be "based on" a book), while aggregation would appear to mean compiled
or collected works (anthologies).  A program linking to another
program is not "based on" that other program in that sense regardless
of the mechanism of linking any more than an anthology is based on the
pieces published therein.  Whether the Geminii table engine would be a
derivative work of MySQL is a question that I don't think the
jurisprudence is clear on (IANAL again).  In the most simplistic of
approaches, Nusphere would be safe.  (It gets complicated because I
don't think API's and can be effectively copyrighted, and header files
are too heavily tied to APIs to get much protection in that way--- see
endless discusson on Groklaw during the SCO case on this issue, but at
the same time, if you can show continuity of expression that goes
beyond functional requirements, then you might have a case.)

But the point here is that both of these are cases where reasonable
minds can disagree.  Rick looks at the BSD license and says "well, it
seems to allow me to license this to others under more restrictive
terms if I keep the old copyright notices and license text in tact."
Someone else might say "sublicensing is not mentioned.  Therefore it's
not allowed."  Again with MySQL v. Nusphere, there are questions where
reasonable people can disagree about the intersection of copyright law
and software regardless of how severe Nusphere's violations of social
norms are.  These are the cases I see getting litigated.  I just don't
see how any statistics there tell us anything useful about the
licenses.

Best Wishes,
Chris Travers